STUDIES ON MOUSE LEUKEMIA* III. A COMPARISON OF FOUR LURES OF LEUXEMIA TRANSMITTED BY

STUDIES ON MOUSE LEUKEMIA* III. A COMPARISON OF FOUR LURES OF LEUXEMIA TRANSMITTED BY INOCULATION BY MAURICE N. RICHTER, M.D., AND E. C. MACDOWELL, S.D. (From the Department of Pathology of the College of Physicians and Surgeons, Columbia University, New York, and the Department of Genetics, Carnegie Institution of Waskington, Cold Spring Harbor) (Received for publication, July 21, 1930) In previous communications we have recorded the existence of an inbred strain of mice designated as C58 in which lymphatic leukemia occurs spontaneously with great frequency, and also the transmission of leukemia to normal young mice of the same strain by inoculation with suspensions of minced tissues (1). We also reported the establishment of several lines of experimentally transmitted leukemia, by transfer from one mouse to another (2). These lines have been designated by letters of the alphabet in the order of their first transfers. Elsewhere (3) we reported the results of inoculation of mice which were genetically different, i.e., mice of the susceptible strain C58, a resistant strain (StoLi), the F-1 hybrids of these strains, and the back cross generations. Differences in the results were obtained which were interpreted as due to differences in the genetic constitutions of the hosts. In this paper we report differences in results of inoculation due to differences not in the hosts, but in the materials inoculated. With the establishment of several lines of experimentally transmitted leukemia, we found certain lesions appearing in particular lines with such frequency that they could be considered as characteristic of the lines. The differences between the lines were mainly those of ex- * This investigation was supported by a grant from the Carnegie Corporation and an appropriation for technical assistance from the Research Fund of Columbia University. The inoculation and supervision of an{reals were in the hands of Miss Mary Van S. McCoy and Miss Katherine Hill in New York. 823

824 MOUSE LEUKEMIA. III tent or distribution of lesions, or the consistency of results. We have, for example, observed that typical leukemic blood counts were more common in certain lines than in others; that pleural or peritoneal effusions were, for a long time, confined to certain lines; that when inoculated into hybrids of the back cross generation, Line I showed fairly consistent proportions of susceptible mice, while the proportion susceptible to Line A varied when tested at different times; that the liver and kidneys presented differences in the frequency and degree of infiltration. Four lines (A, E, H, I) have been carried for a sufficient number of transfers to furnish data on the relative frequency of the occurrence of some of these characteristics. We have selected for this purpose six features for comparison. These are the frequency of leukemic blood counts, pleural and peritoneal effusions, infiltration in liver and kidney, and the variability of results. The four lines were carried on simultaneously, although Line A was in the 10th transfer when the first transmission of Line I was made. The inoculations were made into mice taken at random from the closely inbred Strain C58, and animals of the same litter were frequently inoculated with material from different lines. From this genetic uniformity of the mice, it follows that differences between the lines can not be attributed to differences in the mice. Leucocyte Counts We have previously reported (2) that in spontaneous leukemia the leucocyte count is not always increased, and when increased does not always remain so. Similar variations have also been observed in leukemias transmitted by inoculation. The presentation of data on the occurrence of increased leucocyte counts therefore involves certain difficulties, for it has not been possible to examine the blood of each mouse at short intervals. Daily determinations of the leucocyte counts in Line I have shown that significant changes do not occur until within 3 days of death, at wiiich time sudden increases are usually observed. In other lines, marked changes do not occur within this interval. It is obvious, therefore, that to determine the frequency of this change, the counts must be made within this 3 day period. In many instances the blood

M. N. RICHTER AND E. C. MACDOWELL 825 of only one mouse of a group inoculated with the same material was examined within this time, and this was usually the animal used as donor for the next transfer. Donors are killed instead of waiting for death to occur spontaneously, and this is usually done a day before the death of other mice in the same experiment. Terminal drops in the count are avoided by using only the highest count observed for each mouse within 3 days of death. The first eight transfers of each line are omitted because during this period the counts of Line A varied 0 50 Ioo 2ao- Uo ~e ~o 4o0 4So Lc~ce~ c~lnt~ in ~ FIo. 1. Leucocyte counts in 4 lines of inoculated leukemia. All counts made within 3 days of death. Counts of the donors of these lines were: for Line A, 594,000 per cubic millimeter; Line E, 90,000; Line H, 292,000; Line I, 42,000. considerably, and were higher than those which later became characteristic of the line. Limitation of data in this way does not affect the results except in Line A. A comparison of the leucocyte counts thus tabulated (Fig. 1), shows that increases were observed less frequently in Lines A and E than in Lines H and I, and that the increased counts in the former lines were relatively less marked than in the latter. In Line I the counts were invariably increased and were frequently higher than any counts ob-

826 ~OUSE LEUKEMIA. III served in the other lines. The blood counts after inoculation cannot be correlated with the number of leucocytes in the blood of the spontaneous case used as donor for the first transfer of the line. Peritoneal Effusion In the early transfers of Line A several mice were found to have fluid in the peritoneal cavity. In most instances this was a milky fluid tinged with blood, and was often abundant. The number of mice that had this lesion increased so that after the eighth transfer, nearly every mouse dying of leukemia had peritoneal effusion. Specifically, from the 1 lth to the 23rd transfers inclusive, fluid was present in every positive case. In the 28th and subsequent transfers a smaller number of cases with fluid were observed, and the amount was less. Since the 28th transfer the appearance of the lesion has been inconstant and the amount relatively small. The great majority (83 per cent) of the cases in the first 43 transfers had fluid in the peritoneal cavity (Table 1). In Line E the fluid first appeared in the third transfer, and reappeared in every subsequent transfer (but not in every mouse) through the 14th. During the next 10 transfers (40 mice) only five cases with fluid were found. The lesion was present in 27.6 per cent of the cases in the first 24 transfers. In Lines H and I peritoneal effusion has been observed very rarely. There are thus marked differences in the four lines with respect to the occurrence of this lesion. In Line A it was very frequent, in Line E less so, and in Lines H and I relatively rare. None of the four spontaneous cases used as donors for these lines had fluid in the peritoneal cavity. Pleural Effusion As in the case of peritoneal fluid, that in the pleural cavity is milky, but usually with less admixture of blood. This feature has been observed frequently in Lines A (24.7 per cent) and E (35 per cent), but rarely in Lines H (0.7 per cent) and I (1.2 per cent) (Table 1). The occurrence of fluid in the pleural cavity is less constant than in the peritoneal. It is very rarely observed in all the mice of an experi-

M. N. RICHTER AND E. C. MACDOWELL 827 ment, and has varied at different times. In line A only 5 cases were observed in the first 17 transfers (4 per cent), and 80 cases in the next 24 transfers (38.7 per cent). In Line E there were 41 cases with pleural fluid in the first 14 transfers (43.6 per cent), but only 6 cases in the next 10 (15 per cent). The one case with pleural fluid that occurred in Line H was in the 10th transfer, and peritoneal fluid was present in the same case. Pleural effusion was not present in any of the spontaneous cases used as donors. TABLE 1 Occurrence of Peritoneal and Pleural Effusions in Four Lines of Inoculated Leukemia Line No. mice No. transfers Peritoneal effusion No. mice Per cent Pleural effusion No. mice Per cent A E H I 343 134 144 504 43 24 26 37 285 83.0 37 27.6 4 2.8 3 0.6 85 24.7 47 35.0 1 0.7 6 1.2 Infiltration in the Liver Infiltration in the liver varies in extent and position, and an arbitrary method of classification is necessary. In some animals the infiltration is extensive and present to the same degree in all microscopic sections. In such cases it is usually perivascular, with or without extensive intracapillary collections. These cases are classified as having "marked" infiltration. Less marked degrees of infiltration are of different types. In some, the infiltration is widespread, but consists of fewer cells. In others it is not uniform, but may appear as solitary or multiple lymphomata resembling nodular metastases (sometimes seen in the gross specimen) or, more frequently, as perivascular collections appearing particularly around the portal vein near the hilum of the liver. The rest of the liver may be uninvolved, so that sections which do not include the porta hepatis with the larger vessels may not include the areas of infiltration. These several types of involvement are classified as "slight" infiltration.

828 MOUSE LEUKEMIA. III Although examination of a single section does not exclude infiltration if none is found, it has been more convenient to examine but one section in most instances. The results of such examination show such differences in the four lines that conclusions can be drawn regarding the frequency of marked and slight degrees of infiltration. Reference to Fig. 2 shows that in Lines H and I all of the livers examined were markedly infiltrated, as contrasted to Lines A and E in which some were markedly infiltrated, some had slight degrees of infiltration, and others had none in the section examined. The livers of all of the four spontaneous donors had marked infiltration, a common finding in spontaneous cases. Infiltration in the Kidney Degrees of infiltration in the kidney are classified in the same way as in the liver. Line I had marked infiltration with great regularity, but no instances of slight infiltration. Line H had some cases with marked, but more with slight infiltration. In Lines A and E an extremely small number of cases presented infiltration in the kidney, although the perirenal fat was often densely infiltrated. In spontaneous cases the kidney is infiltrated less frequently than the liver. Of the four spontaneous cases used as donors for these lines, infiltration in the kidney was found only in the donor for Line E, and this was of slight grade. Variability The consistency with which characteristic lesions have appeared has varied greatly in the different lines. Some of these variations have involved changes in the characteristic features of the lines, while others consist merely of greater or less uniformity among mice inoculated with the same material. Some of these.changes have already been noted: in Line A, disappearance of marked increases in the leucocyte counts after the 8th transfer; diminution in number of cases with peritoneal effusion after the 28th transfer; variation in proportion of susceptible to non-susceptible back cross animals (3). In Line E, diminution in number of cases with peritoneal or pleural effusion after the 14th transfer. Corresponding changes have not been noted in Lines H or I.

M. N. RICHTER AND E. C. MACDOWELL 829 Variation among mice in the same experiment has also differed in the four lines. In Lines H and I a considerable degree of regularity in result has been observed. Not only have the mice presented the same lesions, but the survival time, after the first few transfers, has varied only within narrow limits. In Lines A and E on the other hand, greater variation has been noted among mice inoculated with the same material, with respect to the occurrence of particular lesions and to the period of survival. The differences between the survival times in Lines A and I are striking (Fig. 3). The transfers in Line I from the 16th to the 50th, TABLE 2 Line Marked infiltration Slight infiltration I No infiltration Frequency of infiltration of liver A 36 64 71 E 30 32 11 H 85 0 0 I 98 0 0 Frequency of infiltration of kidney A 0 1 98 E 1 1 68 tt 20 36 28 I 87 0 10 The proportion of mice with various degrees of infiltration is indicated graphically in Fig. 2. inclusive, involve 150 mice inoculated intraperitoneally, and cover the period from November 20, 1929, to June 8, 1930. During this period the survival time varied but little. With one exception, the mice died between 6 and 9 days after inoculation, the greatest number (47.3 per cent) on the 7th day. The 16th to 50th transfers of Line A involved 182 mice inoculated intraperitoneally. The range in survival time was from 6 to 7.1 days. The majority (85.2 per cent) died from 6 to 22 days after inoculation, but' there was no great preponderance of deaths on any one day. If, instead of comparing the lines during the same

830 MOUSE LEUKEMIA. III transfers, the comparison is made during the same time interval, the general shape of the frequency curve in Line A is unaltered and the marked difference in regularity persists. Variability is also noted in the distribution of negative results in the four lines. In Lines E, H, and I, unsuccessful inoculations were t0o ~ 8O 6O A E LIVEP, 4.0 20 0 o 1oo,/. KI D N EY 6o 2~ 20 I 0 A E H I FIG. 2. Infiltration in liver and kidney. Degree of infiltration indicated as follows: black = marked infiltration; shaded = slight infiltration; white = no infiltration. Proportion of mice involved indicated by height of rectangles. Data for this table are given in Table 2. obtained only in the first 3 transfers. In Line A, however, negative results were obtained in several later transfers, and as recently as the 35th. Furthermore, Line A is the only one in which recoveries have ap-

M. N. RICHTER AND E. C. MACDOWELL 831 parently occurred. During the period in which peritoneal effusion was an outstanding feature of the line, 5 mice that had swelling of the abdomen later recovered. Inasmuch as this abdominal swelling occurred only in the mice inoculated with Line A, and at the same time as in mice that later died with peritoneal effusion, it is believed that the temporary swelling was due to this lesion and that the mice recovered. "~0. o! 70 60 50 40 t I! :I I i I il I t!! 30 20 10... ib... zo " " : " 3o '... 4o::: go... 6o... :... z0 ~rviva,1 tiln~ i~ c1~y~ FIG. 3. Frequency curves of survival times of mice in Lines A and I. Solid curve = Line A. Broken curve = Line I. DISCUSSION The occurrence within a single inbred strain of mice of cases of leukemia presenting differences in the type, extent or distribution of lesions, and the appearance of the same differences within a single line of experimental transfers indicate that the fundamental process underlying the leukemic lesions may lead to varied manifestations.

832 MOUSE LEUKEMIA. III In previous experiments (3) it was found that hybrid mice from the back cross to the non-susceptible strain were not all susceptible, but that among those which were, the type of response depended on the material inoculated and was the same in the hybrids as in mice of Strain C58 inoculated with the same material. That is, susceptible hybrids inoculated with material from Line A presented lesions found in C58 mice inoculated with Line A; those inoculated with Line I had lesions found in C58 mice inoculated with Line I. It is evident that one of the important factors controlling the susceptibility to leukemia transmitted by inoculation is the genetic constitution of the host; yet the particular manifestations presented by susceptible animals depend not on their genetic constitution, but on the nature of the material inoculated. SUMMARY Several lines of lymphatic leukemia in mice, expcrimentaily transmitted by inoculation into hosts of a closely inbred strain, have bccn cstablished and carried on simultaneously. Among the inoculated mice therc ~vcre found diffcrcnt types of response, according to the line of leukemia inoculatcd. The differences consisted mainly in thc extent or distribution of lesions. Although the same line did not always show the same distribution of lesions, there was a distinct tendency for the cases in a linc to present the same characteristics on successive transfers over a considerable period. A comparison of the frequency of the occurrence of certain lesions in four lines of transmission shows that: I. Incrcascs in the total leucocyte counts were less common and less marked in Lines A and E than in Lines H and I. In the lines first mcntioned, normal counts wcrc rclativcly frequent. In Line H normal counts were uncommon, and in Line I were not observed. The highest counts were found in Line I. 2. Pcritoneai effusion was characteristic of Line A for 43 transfers. It was found in Line E, but for a considerably shorter period and in a smaller proportion of cases; it was rare in Lines H and I. 3. Pleural effusion was frequently found in Lines A and E but rarely in Lines H and I. 4. Infiltration in the livcr occurred to a markcd degree in all the

M. N. RICHTER AND E. C. MACDOWELL 833 mice of Lines H and I, but in fewer mice and to a variable degree in Lines A and E. 5. Infiltration in the kidney in Line I was present in 90 per cent of the mice, and was of marked grade. In line H it was present in 67 per cent of the mice (24 per cent of marked grade) but in Lines A and E it was rare. 6. A considerable amount of variability in results occurred in Lines A and E, whereas Lines H and I remained remarkably constant. The lesions characteristic of a line were not necessarily those present in the spontaneous case from which the first transfer of the line was made. As the mice used for inoculation were genetically uniform, the differences between the lines are not due to genetic differences in the hosts, but to differences in the materials inoculated. REFERENCES 1. Richter, M. N., and MacDowell, E. C., Proc. Soc. Exper. Biol. and Med., 1929, 261 362. 2. Richter, M. N., and MacDowell, E. C., Jr. Exper. Med., 1930, 51,659. 3. MacDowell, E. C., and Richter, M. N., Jr. Cancer Res., 1930, 14, 434.